• Date:  2025-05-26 (월요일) 16:00 ~ 17:00 
• Place:  27동 116호
• Speaker:  이규환 (University of Connecticut)
• Title:  Learning Euler Factors of Elliptic Curves with Transformers
• Abstract: A transformer is a neural network architecture widely adopted in models such as GPTs. We apply transformer models to predict Frobenius traces $a_p$ of elliptic curves, given other traces $a_q$. Our experiments reveal that these models achieve high accuracy, even without relying on explicit number-theoretic tools. We also present preliminary findings on model interpretability.